U.S. patent number 10,882,186 [Application Number 15/976,853] was granted by the patent office on 2021-01-05 for method for efficient operation of mobile robotic devices.
This patent grant is currently assigned to AI Incorporated. The grantee listed for this patent is Ali Ebrahimi Afrouzi, Chen Zhang. Invention is credited to Ali Ebrahimi Afrouzi, Chen Zhang.
United States Patent |
10,882,186 |
Ebrahimi Afrouzi , et
al. |
January 5, 2021 |
Method for efficient operation of mobile robotic devices
Abstract
A method for efficient navigational and work duty planning for
mobile robotic devices. A mobile robotic device will autonomously
create a plan for navigation and work duty functions based on data
compiled regarding various considerations in the work environment.
These factors include what type of work surface is being operated
on, whether dynamic obstacles are present in the work environment
or not and the like factors.
Inventors: |
Ebrahimi Afrouzi; Ali (San
Jose, CA), Zhang; Chen (Dublin, CA) |
Applicant: |
Name |
City |
State |
Country |
Type |
Ebrahimi Afrouzi; Ali
Zhang; Chen |
San Jose
Dublin |
CA
CA |
US
US |
|
|
Assignee: |
AI Incorporated (Toronto,
CA)
|
Family
ID: |
74044916 |
Appl.
No.: |
15/976,853 |
Filed: |
May 10, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
62505060 |
May 11, 2017 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
B25J
11/0085 (20130101); G05D 1/0217 (20130101); B25J
9/1666 (20130101); G05D 1/0221 (20130101); G05D
2201/0203 (20130101); Y10S 901/01 (20130101); Y10S
901/02 (20130101) |
Current International
Class: |
B25J
9/16 (20060101); B25J 11/00 (20060101) |
Field of
Search: |
;700/245,255 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Figueroa; Jaime
Parent Case Text
CROSS-REFERENCE TO RELATED APPLICATIONS
This application claims the benefit of provisional application Ser.
No. 62/505,060, filed May 11, 2017 by the present inventor.
Claims
We claim:
1. A method for a mobile robotic cleaning devices to autonomously
plan work duties in a work environment in a most efficient manner
possible, the method comprising: accessing historical data stored
from prior work cycles, the historical data being indicative of a
type of work surface and a presence or absence of dynamic obstacles
in a given work area; processing probabilities based on the
historical data as to a most efficient navigational route and
operational work duties for the mobile robotic cleaning device to
partake in with one or more processors of the mobile robotic
cleaning device; enacting the most efficient navigational route and
operational work duties by the mobile robotic cleaning device;
capturing new data by one or more sensors of the mobile robotic
cleaning device while enacting the navigational route and
operational work duties within the work environment; processing new
data with the one or more processors of the mobile robotic cleaning
device; assessing a penalty score for a work area within the work
environment when a dynamic obstacle is detected in the work area
with the one or more processors; and enacting a new most efficient
navigational route and operational work duties by the mobile
robotic cleaning device when discrepancies exist between the new
data and the historical data.
2. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device autonomously determines that the
most efficient cleaning manner is to clean one work surface type at
a time before cleaning a different work surface type.
3. The method of claim 2 wherein the mobile robotic cleaning device
thereafter cleans one work surface type at a time completely before
cleaning a second work surface type.
4. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device monitors the presence or absence of
dynamic obstacles in the work environment.
5. The method of claim 4 wherein when a dynamic obstacle is
detected, the one or more processors of the mobile robotic cleaning
device collects data regarding the location this encounter occurred
at, the time this encounter took place, the date of the encounter,
and type of dynamic obstacle encountered.
6. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device autonomously determines to perform
work duties in a different work area upon detecting the presence of
a dynamic obstacle in a given work area.
7. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device autonomously determines based on
processed probabilities that it is likely to encounter a dynamic
obstacle in a work area and should therefore perform work duties in
a different area where a presence of a dynamic obstacle is
unlikely.
8. The method of claim 7 wherein after cleaning the work area the
mobile robotic cleaning device thereafter checks other work areas
where there have been a historical presence of a dynamic obstacle
to determine when to clean those work areas.
9. The method of claim 1 wherein when a penalty score is assessed
for the work area, the mobile robotic device will not conduct work
duties or navigate to that area until the penalty score has reached
a predetermined threshold.
10. The method of claim 1 wherein the penalty score assessed for
the work area decreases at a fixed rate over a fixed time
interval.
11. The method of claim 1 wherein the penalty score assessed for
the work area decreases every time the mobile robotic device
performs work duties in a different work area.
12. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device creates an internal map of the work
environment with specific markers on the map marking the location
of different work surface types and dynamic obstacles detected
during operation.
13. The method of claim 1 wherein the new data consists of the
mobile robotic cleaning device autonomously learning the boundaries
or edges between different work surface types during operation.
14. The method of claim 13 wherein the one or more processors of
the mobile robotic cleaning device learns the boundaries or edges
between different work surface types from other mobile robotic
devices.
15. The method of claim 1 wherein the layout of an entire work area
contributes to the consideration of a mobile robotic cleaning
device determining what constitutes the most efficient navigational
route and operational work duties.
16. The method of claim 1 wherein the amount of battery power left
is a contributing factor to the one or more processors of a mobile
robotic cleaning device determining what constitutes the most
efficient navigational route and operational work duties.
17. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device learns of a presence of a current
dynamic obstacle in a work area from data shared by another mobile
robotic device.
18. The method of claim 1 wherein the one or more processors of the
mobile robotic cleaning device learns of a historical presence of a
dynamic obstacle in a work area from historical data shared by
another mobile robotic device.
19. The method of claim 1 wherein the mobile robotic device is a
mobile robotic device other than a mobile robotic cleaning
device.
20. The method of claim 1 wherein the penalty score quantifies the
preference of cleaning the work area, wherein the preference is
based on at least the presence of dynamic obstacles within the work
area.
Description
FIELD OF INVENTION
The present invention relates to mobile robotic devices generally
and more specifically to how mobile robotic devices operate in a
work environment.
BACKGROUND
Sensory detection is important for mobile robotic devices. Mobile
robotic devices must be able to detect obstacles in order to avoid
running into them resulting in damage to the obstacle or the device
itself. Several inventions exist to demonstrate obstacle detection
and obstacle avoidance. Additionally, changes in surface type that
a mobile robotic device can impact the functions for the robot. For
example, a robot may not be able to operate effectively on a
particular type of surface. Alternatively, surface types may impact
the functions a robot can partake in. However, inefficiencies
remain. A need exists for mobile robotic devices to be able to
autonomously plan the most efficient work operational plan through
a work environment that incorporates considerations regarding the
status of the work environment itself.
SUMMARY
The following present a simplified summary of some embodiments of
the invention in order to provide a basic understanding of the
invention. This summary is not an extensive overview of the
invention. It is not intended to identify key/critical elements of
the invention or to delineate the scope of the invention. Its sole
purpose is to present some embodiments of the invention in a
simplified form as a prelude to the more detailed description that
is presented below.
The present invention introduces a method for mobile robotic
devices to autonomously make decisions regarding the most efficient
navigational and work duty plan for a work environment.
Considerations regarding dynamic obstacles, the type of work
surfaces present, and other relevant considerations are factored
into this decision making. The mobile robotic device will process
probabilities and all collected data regarding the work environment
in order to create the most efficient plan of operation for a work
environment.
BRIEF DESCRIPTION OF DRAWINGS
Non-limiting and non-exhaustive features of the present invention
are described with reference to the following figures, wherein like
reference numerals refer to like parts throughout the various
figures.
FIG. 1 illustrates the process by which a mobile robotic device
detects a boundary edge between multiple work surface types and how
it deals with the data collected.
FIG. 2 illustrates an example of a mobile robotic device learning
the boundary between two different work surface types.
FIG. 3 illustrates an example of a mobile robotic device learning
the boundary between two different work surface types.
FIG. 4 illustrates an example of a mobile robotic device learning
the boundary between two different work surface types.
FIG. 5 illustrates how a mobile robotic device detects a dynamic
obstacle in the work environment and the response process to such a
detection.
DETAILED DESCRIPTION OF THE INVENTION
The present invention will now be described in detail with
reference to a few embodiments thereof as illustrated in the
accompanying drawings. In the following description, numerous
specific details are set forth in order to provide a thorough
understanding of the present invention. It will be apparent,
however, to one skilled in the art, that the present invention may
be practiced without some or all of these specific details. In
other instances, well known process steps and/or structures have
not been described in detail in order to not unnecessarily obscure
the present invention.
In still other instances, specific numeric references such as
"first material," may be made. However, the specific numeric
reference should not be interpreted as a literal sequential order
but rather interpreted that the "first material" is different than
a "second material." Thus, the specific details set forth are
merely exemplary. The specific details set forth are merely
exemplary. The specific details may be varied from and still be
contemplated to be within the spirit and scope of the present
disclosure. The term "coupled" is defined as meaning connected
either directly to the component or indirectly to the component
through another component. Further, as used herein, the terms
"about," "approximately," or "substantially" for any numerical
values or ranges indicate a suitable dimensional tolerance that
allows the part or collection of components to function for its
intended purpose as described herein.
The term "certain embodiments", "an embodiment", "embodiment",
"embodiments", "the embodiment", "the embodiments", "one or more
embodiments", "some embodiments", and "one embodiment" mean one or
more (but not all) embodiments unless expressly specified
otherwise. The terms "including", "comprising", "having" and
variations thereof mean "including but not limited to", unless
expressly specified otherwise. The enumerated listing of items does
not imply that any or all of the items are mutually exclusive,
unless expressly specified otherwise. The terms "a", "an" and "the"
mean "one or more", unless expressly specified otherwise.
The term "dynamic obstacle" shall hereinafter mean obstacles in a
work environment that were not present in a prior operating cycle
but are now present in a current operating cycle as well as
obstacles that are present in a current operating cycle that were
not present previously. Further dynamic obstacle will be intended
to mean any and all obstacles that are moving at the time of
operation including but not limited to pets, humans, other mobile
robotic devices and the like. Additionally, obstacles which are
invisible to the robot will also be included as dynamic obstacles
including but not limited to obstacles that appear or disappear in
regards to the mobile robotic device's sensors such as infrared,
depth camera or any other sensor type included with the robot and
the like.
A method for decision making for mobile robotic devices is
disclosed. During operation, utilizing machine learning, a mobile
robotic device will process probabilities in autonomously
determining where the mobile robotic device should operate as well
as what operations should be partaken in. Information pertaining to
work surface types as well as the presence or absence of dynamic
obstacles in work areas are some, but not all of the considerations
utilized in making these decisions. Utilizing data from prior work
cycles a mobile robotic device will come up with a navigational
plan as well as a plan for work duties to be conducted. The mobile
robotic device's plan will factor in various considerations in
order to come up with the most efficient navigational and work plan
for conducting work duties. While performing work duties, sensors
on the mobile robotic device will collect new data. If the data
requires the mobile robotic to alter its work and navigational
plan, it will do so and store the new data for future use.
In the present invention dynamic obstacles and floor type are
detected by the mobile robotic device during work operation. As the
mobile robotic device operates in the work environment, sensors
arranged on the mobile robotic device will collect information such
as what type of work surface is being operated on. For instance,
for mobile robotic devices which clean, for example a mobile
robotic cleaning device, what type of work surface is being
operated on can be quite important. For example, if a room has a
thick pile rug as well as wood flooring, then this information is
important for the operation of the mobile robotic cleaning device.
In such a situation the presence of two differing work surfaces may
make the operation of the mobile robotic cleaning device difficult
as it has to adjust itself when it transitions from the taller
elevation rug to the lower elevation wood floor. Additionally, the
type of cleaning in such a situation may be different. For example,
the cleaning on rug versus the cleaning on a wood floor may require
different functions and may even elevate the pressure on the robot.
For example, when a mobile robotic cleaning device operates on a
rug, it may require a higher suctioning power or the brush may need
to spin faster than when it is cleaning on a wood floor. Since
debris can become stuck deep in the rug higher suctioning power may
be necessary over that of a wood floor. Alternatively, a mobile
robotic cleaning device that can vacuum as well as mop would result
in different functions based on the floor type as the mobile
robotic cleaning device is not going to mop the carpet.
Additionally, a user may not wish for a mobile robotic device to
operate on certain floor types. Take for example a mobile robotic
device that enters the home to deliver packages. A user may wish
for such a mobile robotic device to only operate on a tiled surface
but not that of a carpeted floor as the mobile robotic device would
potentially track in dirt from outside and tile is easier to clean
than carpet.
In embodiments, as the mobile robotic device traverses the work
environment it will take note of what type of work surfaces are in
the work environment it is operating in and where those work
surfaces are located. The device will access this information
during a live work cycle in order to assist with decision making on
the part of the mobile robotic device. In order to maximize the
efficiency of a work operation for the mobile robotic device, the
mobile robotic device will autonomously determine probabilities
based on data it has collected over time regarding the work surface
types in rooms and what rooms it should operate in in order to
maximize operational efficiency. For example, for a mobile robotic
cleaning device, the mobile robotic cleaning device may determine
that it is more efficient to clean the bedrooms first as they
consist of a single work surface type, for example a carpeted
surface, while other rooms contain multiple types of work surfaces
and that transitioning between multiple surface types would be
burdensome and time consuming for the mobile robotic cleaning
device. Further, in a similar situation, a mobile robotic cleaning
device may determine that it should clean a hallway first as it
consists of a wood floor and the wood floor can be cleaned quicker
than rooms where the floor consists solely of carpet. Rooms with
multiple floor types are burdensome for the mobile robotic device.
For example, for devices such as mobile robotic cleaning devices,
operating in a room with multiple floor surface types is burdensome
for the robot because it must transition itself between different
surface types resulting in different functionality, physically
putting pressure on the robot, as well as lowering battery
efficiency. As it traverses each edge or boundary from one surface
type to another a mobile robotic cleaning device may need to switch
the amount of power it is utilizing to clean the different surface
types for example by increasing suction power or brush rotation
speed thereby wasting time and energy each time it transitions.
Further, it may impact the angle of the mobile robotic cleaning
device resulting in the brush and suction section missing contact
with the work surface for a period of time as it will be at an
awkward angle during this transition period. It is preferable for a
mobile robotic cleaning device to finish cleaning a single surface
type at a time before moving onto a different work surface type.
With each cycle, the mobile robotic device will compile such data
in order to assist with decision making during the next operational
cycle.
In some embodiments, the mobile robotic device may assess that it
should autonomously operate on a single work surface type at a time
even if multiple surface types are present in the same work area.
For example, for a mobile robotic cleaning device, the mobile
robotic cleaning device may determine that it is preferable to
operate on rug before transferring to a wood floor. Another example
could be a mobile robotic device that carries items. The mobile
robotic device may determine that it is preferable to operate on a
single surface type if it is carrying delicate materials as
alternating between different surface types may make the items it
is carrying unstable and susceptible to breaking when the mobile
robotic device encounters the elevation change between the
different surface types at their boundary or edge. Whatever the
situation, during the course of normal operation a mobile robotic
device will learn the boundaries and edges between different work
surface types. This process is demonstrated in FIG. 1. The mobile
robotic device will learn these boundaries over a period of time as
it makes multiple operational cycles in a given area. For example,
as seen in FIG. 2 if a mobile robotic cleaning device is attempting
to clean a rug 201 completely first in a room that sits on top of a
wood floor 202 it may clean beyond the boundary of the rug and
clean a small amount of the wood floor. It will learn where the
boundary is and store this data. During the next cleaning cycle,
after accessing the data from day 1, the mobile robotic cleaning
device may once again clean beyond the boundary but the error
margin may be smaller than on the first day. On a third day the
mobile robotic cleaning device may clean too short of the boundary
between the rug and wood floor, see FIG. 3. Ultimately all of this
data will be compiled. On a fourth day the mobile robotic device
may then get the boundary right and clean only the rug area but not
on the wood floor see FIG. 4, or the mobile robotic device may
completely clean one work surface type before transferring over to
clean the second work surface type. The mobile robotic device will
compile data from each cleaning cycle in order to assess
probabilities about where it should clean. The days for a mobile
robotic device to learn the boundaries and edges of work surfaces
are not intended to be predetermined. They are to be machine
learned and the mobile robotic device will learn them at the mobile
robotic device's own individual pace.
Ultimately, a mobile robotic device will have a map it has
autonomously generated based on the collected data of the work
environment it has compiled over time. It will create and update
the map with new data after each additional work cycle.
In some embodiments, in an environment where multiple mobile
robotic devices are present, the mobile robotic devices may share
information pertaining to the specific location, type of work
surface, as well as internal maps generated, and historical
operational data with each other. In some embodiments, the mobile
robotic device will not need to learn the boundaries of work
surfaces on its own as it will learn this information from the data
shared from other mobile robotic devices. In some embodiments, the
mobile robotic device will learn where the boundary edges are
between surface types but will still need to confirm these during
normal working operation. In alternative embodiments, the mobile
robotic device will still need to autonomously learn the boundaries
between varying work surface types on its own even after these
boundaries have been shared by other mobile robotic devices.
In some embodiments a user may use an APP paired with the mobile
robotic device. The user can thereafter diagram in a given work
area roughly where different work surface types are located in
order to assist the mobile robotic device with detecting such work
surfaces. The mobile robotic device will thereafter anticipate a
work surface change before it encounters it for that location in
the room. However, the mobile robotic device will still need to
autonomously learn the specific boundaries between varying work
surface types.
Another aspect of the mobile robotic device operating at peak
efficiency deals with interaction with dynamic obstacles. A mobile
robotic device may act as a nuisance or safety hazard to dynamic
obstacles such as a human, pet and may interfere with the operation
of other mobile robotic devices. In such a situation it is
undesirable for the mobile robotic device to operate in the same
vicinity at such a time. Additionally, dynamic obstacles can
interfere with the operation on the part of the mobile robotic
device itself. Therefore, in order to operate at peak efficiency,
the mobile robotic device will attempt to limit interaction with
dynamic obstacles when in operation if the interaction will cause
issues.
In embodiments, the device will mark the locations of obstacles
including dynamic obstacles it has encountered in an internally
generated map. Each time an obstacle is encountered the mobile
robotic device will mark the location and nature of the obstacle on
the internal map. Over time if an obstacle repeatedly appears the
mobile robotic device may mark that obstacle as a permanent
obstacle in the environment. In such a situation, the device will
expect that obstacle to be present during work operation. In a
situation where said obstacle is not present, the mobile robotic
device may take note of the missing obstacle. If this repeatedly
occurs then the mobile robotic device may mark the missing obstacle
as a dynamic obstacle. Alternatively, in such a situation the
mobile robotic device may expect the area to not have an obstacle
at all and if one appears will mark the obstacle as dynamic. If
such an obstacle appears on some occasions but not others the
mobile robotic device may mark that obstacle as a dynamic obstacle.
In some embodiments the mobile robotic device may mark obstacles
that are unexpected, have disappeared or have unexpectedly appeared
as dynamic obstacles.
In some embodiments, a mobile robotic device may encounter
obstacles it has not encountered before in a location but which is
close to where it routinely encounters obstacles. An example is a
dining room chair. If the chair has been moved close from where it
is usually located but not a large distance, for example a foot or
so, then the mobile robotic device may anticipate that this is
merely the same chair that has been slightly moved out of place. In
such a situation the distance is not a predetermined distance but
is up to the mobile robotic device to autonomously determine
whether or not to label such an obstacle as dynamic or not.
In embodiments, when the mobile robotic device encounters dynamic
obstacles that are moving, such as other mobile robotic devices,
pets, humans and the like, the mobile robotic device will collect
this data. The data collected may include but is not limited to
what time this occurred, the date this occurred, where this
occurred such as what room this occurred in, what was likely
encountered such as whether this was a mobile robotic device,
human, pet, piece of furniture or the like, and any other relevant
information that the robot can utilize for cataloguing the
encounter.
In embodiments, when the mobile robotic device encounters an
obstacle that is invisible to it, for example an obstacle that the
sensors on the mobile robotic device were not able to detect, the
mobile robotic device may catalog that obstacle as a dynamic
obstacle.
In some embodiments, in areas with multiple historical occurrences
of dynamic obstacles, the mobile robotic device will expect such a
location to be a high volume traffic location. In these situations,
the mobile robotic device will have catalogued data in regards to
when these instances occurred. For example, at 6 pm a mobile
robotic device may encounter multiple humans in a kitchen or dining
area on multiple occasions. In such a situation the mobile robotic
device would internally note these encounters including information
regarding the fact that it encountered multiple dynamic obstacles,
that this occurred in the kitchen or dining area, that this
occurred at 6 pm, and that this occurred on multiple occasions. The
mobile robotic device would thereafter be able to rely on this
information for future work cycles.
In embodiments, the presence of dynamic obstacles over a period of
time is measured. Utilizing machine learning, the mobile robotic
device processes probabilities and a number between 1 and 10 is
assessed as a penalty score to a given room or work area to aid in
determining what area is preferable to clean first. For example, if
the mobile robotic device encounters a dynamic obstacle such as
multiple individuals in a room at the same time on a given day, it
may assign a high penalty score for that room. At such a time the
mobile robotic device will then leave this room and will attempt to
operate job duties in a different room before returning to complete
job duties in the room it previously assessed a high score for. On
its way to or from the room with the high score, the mobile robotic
device may encounter other rooms as well. When this occurs the
mobile robotic device may check on these other rooms. If the mobile
robotic device determines to check these other rooms, it will enter
and assess whether or not it is an optimal time to perform job
duties in these rooms based on the presence of dynamic obstacles.
If a dynamic obstacle is encountered a penalty score may be
assessed for this particular time on the particular day for this
particular room. The mobile robotic device may then determine
whether or not to conduct a work duty in that room or whether to
move on to another room. Over a predetermined amount of time, the
penalty scores for rooms that have been assessed a penalty will be
reduced at a fixed rate over a fixed time interval. Additionally,
the penalty scores for rooms that have been assessed a penalty will
be reduced every time the mobile robotic device performs a work
duty in a different room. When the mobile robotic device's penalty
score for a particular room is lowered and reaches a predetermined
low enough threshold the mobile robotic device will return to the
room with the lowered score and attempt to perform work duties
again in that room. The mobile robotic device may encounter other
rooms on the way to the room with the lowered score and may attempt
to perform job duties in those rooms first. If when the robot
returns to the lowered score room it encounters enough dynamic
obstacles that it believes it must leave the room again, the mobile
robotic device will once again assess a penalty score for this room
and continue on to additional rooms. FIG. 5 demonstrates the
process by which a mobile robotic device deals with dynamic
obstacles.
In some embodiments, the mobile robotic device may program itself
to avoid areas where the historical data indicates the existence of
a high volume of traffic or high likelihood of dynamic obstacles at
the times it has noted have a tendency to have a high traffic
volume or high likelihood of a dynamic obstacle being present. For
example, for a mobile robotic cleaning device in a home, for a
cleaning cycle running between 5 and 7 pm, the mobile robotic
device may avoid the kitchen and dining areas around 6 pm as it
will expect to encounter humans. Instead, at 6 pm the mobile
robotic cleaning device may deem it better to clean an office,
bedroom, or other low traffic volume areas before returning to
clean the kitchen and dining areas at 7 pm when the humans are no
longer present.
In some embodiments, a user may be able to let the mobile robotic
device know via an APP paired with the mobile robotic device where
items in a room are located. For example, for a dining area a user
can mark where the table and chairs are located. In such a
situation, when entering a given room, the mobile robotic device
will anticipate such obstacles to be in the room and will not mark
them as dynamic obstacles even if it has never encountered them
before.
In some embodiments, a user may be able to let the mobile robotic
device know via an APP paired with the mobile robotic device that
humans, other mobile robotic devices, pets, or other dynamic
obstacles are present in the area.
In some embodiments, in an environment where multiple mobile
robotic devices are present, the mobile robotic devices may share
information pertaining to the specific location, presence of
dynamic obstacles, internal maps generated, and historical
operational data with each other.
In embodiments, the work area map regarding locations of rooms and
layout will play an important factor in regards to the peak
efficiency for operation in regards to the mobile robotic device.
Based on the location of rooms or map layout the device may
determine that it should operate in particular rooms first as
traveling to distant rooms would be burdensome and take time and
battery life. For example, for a mobile robotic cleaning device, if
it is determined that there is a high probability of a dynamic
obstacle in a home office at a particular time but that there is a
very low likelihood of a dynamic obstacle in bedroom 3, then the
mobile robotic cleaning device may determine that it should clean
bedroom 3. However, in a map layout, bedroom 3 may be several rooms
away. Therefore in the interest of operating at peak efficiency,
the mobile robotic cleaning device will attempt to clean the
hallway and bedrooms 1 and 2 which are on the way to bedroom 3. In
an alternative situation, if the mobile robotic cleaning device
determines that on the way to bedroom 3 that the hallway and
bedroom 1 have a low probability of a dynamic obstacle but that
bedroom 2 has a medium probability of a dynamic obstacle then it
may attempt to clean the hallway and bedroom 1 before checking on
bedroom 2 to see if there is a dynamic obstacle present on the way
to bedroom 3. In the same scenario, the mobile robotic device may
simply determine to clean the hallway and bedroom 1 before skipping
bedroom 2 on the way to cleaning bedroom 3. Once bedroom 3 has been
cleaned the mobile robotic device may come back to check to see if
bedroom 2 should be cleaned.
In embodiments, in order to maximize efficiency, the amount of
battery power left may play a factor into what the mobile robotic
device autonomously determines it should perform with regard to
work tasks and navigation patterns. For example, for a mobile
robotic cleaning device, if only one quarter battery life remains,
the mobile robotic cleaning device may utilize this information for
making determinations as to what work tasks should be performed.
For example, if a mobile robotic cleaning device has the option of
cleaning a hallway with a wood surface, or cleaning a large living
room with multiple surface types, the mobile robotic cleaning
device may determine based on battery power left that it would be
optimal to completely clean the hallway and then recharge itself as
the living room with multiple surface types would drain the battery
to the point that the mobile robotic device would not be able to
completely clean the room before charging. In some embodiments, the
location of the mobile robotic device's docking or charging station
may play a factor in the mobile robotic device's autonomous task
decision making process. For example, if a mobile robotic device
has multiple tasks assigned but it cannot complete all of them
before it must recharge itself, the mobile robotic device may deem
that the best actions to be completed are those nearest to the
mobile robotic device's docking or charging station rather than
those which are farther away.
In some embodiments, one factor which may impact the autonomous
decision making on the part of a mobile robotic device is when
tasks were last completed in a given room. For example, for a
mobile robotic cleaning device, the fact that some rooms were
cleaned several weeks ago while others were cleaned only a week ago
may impact the decision making process for the mobile robotic
device. In such a situation, a mobile robotic device may factor in
and prioritize those work duties which have not been completed in
quite some time over those that have been more recently
completed.
In some embodiments a mobile robotic device may factor in the
amount of activity that takes place in a given room. For example,
for a mobile robotic cleaning device, if human activity takes place
more often in the living room than in a home office, the mobile
robotic cleaning device may prioritize cleaning the living room on
a more consistent basis as it will get dirty more quickly.
Additionally, in some embodiments, for a mobile robotic cleaning
device, historical data pertaining to the amount of debris
historically collected in various locations may be considered into
the prioritization of cleaning particular rooms over others.
Additionally, once this data has been catalogued it may be made
available to a user. A user, may thereafter be able to select what
areas they would like a mobile robotic device to operate in and
what actions a user would like the device to partake in.
With all of this information compiled over time, the mobile robotic
device will work to come up with the most efficient navigational
and work duty plan autonomously. The mobile robotic device will
determine what rooms to operate in and what tasks to perform. For
example, for a mobile robotic cleaning device the compiled data
will impact what rooms to clean, how to navigate to them, what type
of cleaning is to take place and when to clean them. The mobile
robotic device will autonomously create a cleaning scheduling.
Utilizing machine learning to process various probabilities, the
scheduling will focus on achieving the most efficient manner for
work tasks possible for the mobile robotic device. The data
compiled over time will factor in various considerations regarding
the work environment including work surface types, room locations,
the presence or absence of dynamic obstacles, battery life and
other relevant and the like data. The mobile robotic device will
add data from each cleaning cycle. If new data emerges from a given
cleaning cycle that is different from prior historical data, this
may impact how the mobile robotic device operates in future work
cycles.
While this invention has been described in terms of several
embodiments, there are alterations, permutations, and equivalents,
which fall within the scope of this invention. It should also be
noted that there are many alternative ways of implementing the
methods, devices and apparatuses of the present invention.
Furthermore, unless explicitly stated, any method embodiments
described herein are not constrained to a particular order or
sequence. Further the Abstract is provided herein for convenience
and should not be employed to construe or limit the overall
invention, which is expressed in the claims. It is therefore
intended that the following appended claims to be interpreted as
including all such alterations, permutations, and equivalents as
fall within the true spirit and scope of the present invention.
* * * * *